Method · The Trigger

Methodology

How the record was assembled.

Confidence system

HIGH

CDC WISQARS data (2019–2024). FBI Uniform Crime Reports (1960–2024). Mother Jones mass shooting database (1982–2024). These are primary sources with established methodology and public audit trails.

CANDIDATE

Historical homicide rate (1900–1960). Known benchmark years (1900, 1910, 1920, etc.) are drawn from published vital statistics. Intervening years are linearly interpolated. The overall arc is directionally correct; individual years may vary by ±1.0 per 100k.

SPECULATIVE

Synthetic state law-strength index. Rankings are composites derived from Giffords Law Center, Everytown for Gun Safety, and RAND Corporation assessments. They are not primary counts of statutes. The underlying Siegel State Firearm Laws Database (134 provisions, 50 states, 1991–2016) was not directly available for this build.

Data sources

CDC WISQARS Firearm homicide, suicide, and total death counts and rates. State-level data 2019–2024. National data with breakdown by intent. FBI UCR Uniform Crime Reports, historical homicide rates 1960–2024. Supplementary Homicide Reports for weapon breakdowns. Mother Jones Mass Shooting Database, 1982–2024. Tracks incidents with 3+ fatalities (updated to 4+ for recent years). Includes location, weapon type, demographics. Gun Violence Archive Daily incident tracking, 2013–2018 in our data. Broader definition than Mother Jones (4+ shot, not necessarily killed). Historical rates U.S. vital statistics, published DOJ/BJS compilations, and academic literature (Eckberg 1995, Monkkonen 2001) for 1900–1960 benchmark years. Giffords / Everytown / RAND State-level gun law rankings used to construct synthetic restrictiveness scores. The RAND State Firearm Law Database tracks 134 provisions across 14 categories.

Pipeline

01_rate_analysis.py — Century-long homicide rate with era annotations

02_law_analysis.py — State law density index and divergence computation

03_mass_shootings.py — Mother Jones data: per-decade, AWB analysis

04_breakdown.py — Gun death breakdown by type (homicide, suicide, accident)

05_export.py — Combine all outputs into meta.json

Limitations

Pre-1960 homicide rates are estimates. National vital statistics before 1933 covered only the "registration states." Rates for 1900–1932 are reconstructed from limited data and academic estimates. The shape of the curve is reliable; individual year values are approximate.

State law scores are synthetic. Without direct access to the Siegel database for this build, we used composite rankings from Giffords, Everytown, and RAND. These agree on the extremes (California vs Mississippi) but may disagree on middle-ranked states.

Correlation is not causation. The divergence between high-law and low-law states is real. Whether the laws cause the difference, or whether states with lower violence are more willing to pass laws, is a question this report presents but does not answer.

Mass shooting definitions vary. Mother Jones uses a stricter definition (indiscriminate public attacks) than the Gun Violence Archive (4+ shot in a single incident). This report uses Mother Jones for the timeline and GVA for incident counts where noted.

Suicide data before 2019 is estimated. The breakdown chart uses published CDC rates for 2019–2024 and historical estimates from academic literature for earlier years. The crossover point (suicide exceeding homicide in the 1980s) is well-documented in the literature.

Acknowledgments

This report draws on data assembled and maintained by the Centers for Disease Control and Prevention (CDC WISQARS), the Federal Bureau of Investigation (Uniform Crime Reports), Mother Jones magazine (Mass Shooting Database), the Gun Violence Archive, the RAND Corporation (State Firearm Law Database), Giffords Law Center, and Everytown for Gun Safety.